Smart glove Berbasis Sensor Flex dan IMU untuk Penerjemah Bahasa Isyarat Indonesia (SIBI) dengan Integrasi Sensor EMG

Fecilia, Katherina Agatha (2026) Smart glove Berbasis Sensor Flex dan IMU untuk Penerjemah Bahasa Isyarat Indonesia (SIBI) dengan Integrasi Sensor EMG. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023221028-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Komunikasi antara penyandang tunarungu dan masyarakat umum masih menghadapi kendala akibat terbatasnya pemahaman terhadap Sistem Isyarat Bahasa Indonesia (SIBI). Smart glove berbasis sensor telah banyak dikembangkan, namun masih mengalami kesulitan dalam membedakan gestur dengan bentuk tangan yang serupa. Penelitian ini mengembangkan smart glove multimodal yang mengintegrasikan lima sensor flex, dua sensor Force Sensitive Resistor (FSR), sensor Inertial Measurement Unit (IMU) BNO055, dan instrumentasi elektromiografi (EMG) MyoWare 2.0 berbasis mikrokontroler STM32F103C8T6. Sinyal time-domain dari seluruh sensor diproses menggunakan Discrete Wavelet Transform (DWT) dengan wavelet Haar, kemudian diklasifikasikan menggunakan Linear Support Vector Machine (Linear SVM). Hasil pengujian menunjukkan bahwa konfigurasi DWT level 7 dengan EMG memberikan performa terbaik dengan akurasi 88,86%, precision 88,94%, recall 88,59%, dan F1-score 88,45%. Konfigurasi DWT level 7 dipilih karena merupakan konfigurasi yang mampu mempertahankan performa klasifikasi sekaligus menghasilkan kebutuhan memori yang cukup ringan, yaitu sekitar 58,4 KB Flash dan 12 KB RAM, sehingga dapat diimplementasikan pada mikrokontroler STM32 tanpa memerlukan reduksi dimensi tambahan. Penambahan EMG meningkatkan akurasi klasifikasi offline dari 88,09% menjadi 88,86% pada konfigurasi DWT level 7. Hasil tersebut menunjukkan bahwa kombinasi DWT dan EMG mampu meningkatkan kemampuan sistem dalam mengenali gestur SIBI sekaligus mendukung implementasi pada perangkat embedded.
=====================================================================================================================================
Communication between individuals with hearing impairments and the general public remains challenging due to the limited understanding of the Indonesian Sign System (SIBI). Sensor based smart gloves have been widely developed; however, they still face difficulties in distinguishing gestures with similar hand configurations. This study develops a multimodal smart glove integrating five flex sensors, two Force sensitive resistor (FSR) sensors, a BNO055 Inertial Measurement Unit (IMU), and a MyoWare 2.0 electromyography (EMG) module based on the STM32F103C8T6 microcontroller. Time domain signals from all sensors were processed using the Haar Discrete wavelet transform (DWT) and classified using a Linear Support Vector Machine (Linear SVM). The experimental results show that the DWT level 7 configuration with EMG achieved the best performance, with an accuracy of 88.86%, precision of 88.94%, recall of 88.59%, and an F1-score of 88.45%. This configuration was selected because it maintained competitive classification performance while requiring only approximately 58.4 KB of Flash memory and 12 KB of RAM, enabling implementation on the STM32 microcontroller without additional dimensionality reduction. Furthermore, the inclusion of EMG improved the offline classification accuracy from 88.09% to 88.86%, demonstrating its effectiveness in enhancing SIBI gesture recognition for embedded applications.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	smart glove, Sistem Isyarat Bahasa Indonesia, elektromiografi, Discrete Wavelet Transform, Linear Support Vector Machine, Smart glove, Indonesian Sign Language System, electromyography, Discrete wavelet transform, Linear Support Vector Machine
Subjects:	T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Katherina Agatha Fecilia
Date Deposited:	01 Aug 2026 02:52
Last Modified:	01 Aug 2026 02:52
URI:	http://repository.its.ac.id/id/eprint/141335

Actions (login required)

View Item